The fluid catalytic cracking (FCC) process has become well-established in the petroleum refining industry for converting higher boiling petroleum fractions into lower boiling products, especially gasoline.
In the FCC process, a finely divided solid cracking catalyst promotes cracking reactions. The catalyst is in a finely divided form, typically with a particles of 20-100 microns, with an average of about 60-75 microns. The catalyst acts like a fluid (hence the designation FCC) and circulates in a closed cycle between a cracking zone and a separate regeneration zone.
In the cracking zone, hot catalyst contacts the feed so as to effect the desired cracking reactions and coke up the catalyst. The catalyst is then separated from cracked products which are removed from the cracking reactor for further processing. The coked catalyst is stripped and then regenerated. A good overview of the importance of the FCC process, and its continuous improvement, is reported in Fluid Catalytic Cracking Report, Amos A. Avidan, Michael Edwards and Hartley Owen, in the Jan. 8, 1990 Oil & Gas Journal.
One of the most significant problems remaining is post-riser cracking. As FCC technology has improved, post riser cracking has gone from a trivial problem which was hard to find in a commercial unit to a major problem which can not be ignored.
The problem could be largely ignored when the FCC unit operated with a riser top temperature of 950° F. As catalysts became more active, and feeds heavier, riser top temperatures increased above 1,000° F. in many units. The large volume “reactor”—in which the riser reactor discharged—became a thermal cracker. Large amounts of cracked product vapor spent a significant amount of time in the vapor space above the FCC stripper, and time and high temperature predictably produced thermal cracking.
More details about the problem, and the state of the art methods of minimizing such thermal cracking, are presented in FCC CLOSED CYCLONE SYSTEM ELIMINATES POST-RISER CRACKING, by Amos A. Avidan, Frederick J. Krambec, Hartley Owen, and Paul H. Schipper, presented at the 1990 NPRA Annual Meeting, Mar. 25-27, 1990.
The authors, one of whom is the present inventor, reported the dramatic decrease in thermal cracking which could be achieved using a “closed-cyclone” system to separate cracked products from spent catalyst as they are discharged from the FCC riser reactor. The paper reviewed various “rough cut” separation devices (in U.S. Pat. No. 4,295,961 and U.S. Pat. No. 4,664,888 and U.S. Pat. No. 4,721,603, which are incorporated by reference). While improvements over dense bed cracking units, or riser cracking units with no cyclones, the rough cut devices merely separated spent catalyst from spent products. The rough cut cyclones still allowed a significant amount of hydrocarbons to remain a long time in the reactor vessel, the volume above the stripper.
The authors presented a new type of riser reactor cyclone, a “closed cyclone” which effectively separated cracked products from spent catalyst and quickly removed the cracked products from the reactor vessel. Closed cyclone designs are reported in U.S. Pat. No. 5,055,177, Haddad et al. This design, and the closed cyclone designs of other oil companies, generally did an excellent job of quickly removing from the vessel the cracked vapors recovered via the riser cyclones. These design ignored another problem, which was thermal cracking of stripper vapor.
A significant amount of the cracked vapor product stays with, or is entrained in, or is needed to fluidize, the spent catalyst. Those working in this field concentrated their efforts on the primary product vapor stream, the 90+ mole % of the cracked vapor recovered as a vapor phase out of the reactor vessel. They generally ignored the secondary vapor product, the modest amount of vapor product that was discharged with the spent catalyst from the reactor cyclone systems. This vapor product, plus additional hydrocarbons displaced or desorbed from the spent catalyst by steam stripping, was at a temperature approaching that of the riser reactor. Although small in size, the secondary product could crack thermally, and was severely overcracked even as the primary product was removed relatively unscathed due to thermal cracking.
Thermal cracking depends on time and temperature. The net effect of getting most of the vapor out quickly (to practically eliminate thermal cracking of this material) was significantly offset by large amounts of thermal cracking of the secondary vapor product in the reactor volume above the stripper.
An order of magnitude less hydrocarbon was discharged down the riser cyclone diplegs to be recovered with stripper vapors. The reactor volume stayed the same, so as closed cyclone efficiency increased, the stripper vapor residence time increased dramatically.
Refiners reported dramatic yield benefits. The Avidan et al paper presented at the 1990 NPRA meeting reported a reduction of 40 percent in sulfur-free dry gas make. This was a significant and noteworthy accomplishment, but was actually a combination of two phenomena—an even more dramatic reduction in dry gas make in the primary product and an offsetting significant increase in thermal cracking and attendant dry gas production in the secondary, or stripper vapor product.
Another benefit of closed cyclones reported in this paper was that the production of butadiene is reduced by over 50 percent. Butadiene is a sensitive measure of thermal cracking in FCC, refiners watch it because it is a major contributor to acid consumption in downstream alkylation units.
I wanted to retain the benefits of closed cyclone FCC riser cracking, but avoid the excessive thermal cracking which occurred in the reactor volume. My goal could also be considered as a way to help solve a long standing problem, that post riser thermal cracking in the vessel volume above the FCC stripper.
Four approaches have been proposed or used to reduce thermal cracking in this area: dome steam, riser quench, post-riser steam quench and a stripper cap/snorkel. To me all these are related and are attempts to reduce thermal cracking in a catalytic cracking unit. Each approach has benefits and burdens, and each is reviewed below.
Dome Steam
Refiners have known for years that thermal cracking went on in the “reactor volume” above the stripper. Thus in addition to the four kinds of coke make associated with FCC (catalytic, CCR, Pt function, and cat:oil), refiners have known about “dome coke”—a product of undesired thermal cracking in the dome of the reactor vessel, or the vessel holding the stripper. Unless this part of the vessel is continuously purged with steam, the relatively stagnant regions allow thermal cracking to proceed unabated, which produces large coke deposits which can grow in size, break off, and damage vessel internals. Refiners now solve this problem by adding 500 to 1000 #/hour of steam directly to the dome volume. This purge steam sweeps hydrocarbons out of the dome region. Refiners have practiced this for decades, but it is so well accepted and universally practiced that it is rarely discussed. I mention it to show that refiners are well aware of the problem of thermal cracking in the vessel volume.
While use of dome steam has been practiced for decades, another attempt at suppressing thermal cracking was commercialized in recent years—riser quench.
Riser Quench
Refiners have known that higher riser temperatures are beneficial in cracking heavy feeds in a short time. They have also known that higher temperatures in the riser lead to undesired thermal cracking downstream of the riser in the reactor vessel. A way to have higher temperatures in the riser than in the reactor vessel is to use riser quench. Several approaches to quench have been developed, as reported in U.S. Pat. No. 4,818,372, Mauleon et al and U.S. Pat. No. 5,389,232, Adewuyi et al, which are incorporated by reference. Basically riser quench involves cracking the feed in the base of the riser at a higher than normal temperature and injecting a cooler material, such as light cycle oil, higher up in the riser.
Some refiners quench quickly, within less than a second of residence time in the riser, while some quench higher up or even at the riser outlet. This will reduce post-riser thermal cracking. A somewhat related approach is post-riser quenching, discussed next.
Post-Riser Quenching
U.S. Pat. No. 4,978,440, Krambeck, Dec. 18, 1990, taught injection of water or steam downstream of the FCC riser. This patent is incorporated by reference.
The patentee recognized that thermal cracking occurred in the reactor vessel. Closed cyclone operation increased the residence time of the stripper vapor. The solution, adding steam to reduce the temperature in the vapor phase above the stripper, reduced thermal cracking. It also required adding a large amount of steam to the FCC unit, and this steam tied up a significant portion of the plant volume with water vapor.
Stripper Cap
A fourth approach was isolation of the stripper, a stripper cap and snorkel, as disclosed in U.S. Pat. No. 4,946,656, VENTED STRIPPER SECTION FOR A FLUID CATALYTIC CRACKING UNIT APPARATUS, which is incorporated by reference. They proposed to isolate the stripper with a cap 40 from the interior volume of the vessel 1. This vessel contained the outlet of the riser reactor, the closed cyclone separation system, and the catalyst stripper. The stripper cap was a slant tray, the slope ensured that catalyst falling on the cap would eventually fall down into the stripper. The isolated stripper vapors were removed from the under the stripper cap via a “chimney vent” line 30 passing through holes 29 in the stripper cap. The vent line tied in with the vapor line from the primary cyclone 5 to the secondary cyclone 9.
The patentees recognized the problem—thermal cracking of stripper vapor in closed cyclone FCC operation. Their solution, if implemented, would partially solve the thermal cracking problem while creating other problems.
Their solution would efficiently remove much of the stripper vapor and reduce—but not eliminate—thermal cracking of stripper vapor in the vessel 1. Thus it was at least a partial solution to the problem, with the part left unsolved being the undesired thermal cracking of stripper vapor which would pass through the holes in the cap to the interior of vessel to eventually leave via an annulus in the upper part of vessel 1.
Implementing this solution in commercial FCC units would cause some problems. First the construction and servicing of the unit are greatly complicated by the addition of stripper cap 40. It physically isolate the stripper from everything else, making it harder to inspect, work on, or repair internal stripper hardware. Second, the cap has to be mechanically strong. If any part of it falls off the operation of the catalyst stripper will suffer greatly. The cap has to be segmented, so that it can be fit through manways providing access to vessel 1. The cap will generally be installed around the cyclone diplegs and the riser reactor, so a complex field fabrication procedure will be required.
The support of chimney vent 30 causes significant problems, at least in the embodiment shown in the patent. The vapor inlet to the secondary cyclone, line 23, is supported entirely by the secondary cyclone. This line in turn supports stripper vent 30, while a portion of line 23 must fit loosely over the primary cyclone vapor outlet, line 21. A loose fit is necessary because the stripper cap has so many holes in it that a significant amount of stripper vapor will make its way through the reactor vessel volume and pass through annulus 27.
The mechanical design of such a system is complex and costly. The system must accommodate a significant amount of thermal expansion. Just as the SR71 Blackbird is reported to grow in length almost a foot, as it heats up during supersonic flight, an FCC riser gets longer as it heats up. The riser heats up first, followed by the cyclones, so they do not expand simultaneously. In addition to thermal stress, the FCC riser can bump and even shake at times if something goes off on cat:oil ratio or steam addition in some part of the unit.
An additional concern with the vent cap design shown is possible formation of dome coke, or perhaps “vent cap coke”, in stagnant portions of the cap. Even though steam is added in large amounts to the stripper, and apparently to the stripper volume via line 44, the apex regions of the vent cap will be relatively stagnant and difficult to purge with steam. Coke formation is a distinct possibility. Once coke formation starts, it will continue. It is possible to form large pieces of coke which can fall off and impair stripper operation and may even present a safety hazard during turnarounds. It may be possible to design a vent cap purge which would be as effective as dome steam but this is not a trivial problem, as this area is so difficult to reach. If a small steam line is put in it may not be there in a few years (due to erosion), while if a large line is put in there are problems supporting it and connecting it to a “moving target”, which will be the case due to thermal expansion.
The most troublesome concern is that the design seems to allow a significant amount of hydrocarbon vapor traffic from the stripper up through the reactor vessel into annulus 27. A significant amount of thermal cracking of this material as it passes through the large void volume of the reactor vessel seems likely. The patentees recognized this, and proposed adding more steam via steam spargers 34 and 36 to speed this material on its way. While this will mitigate the damage done to stripper vapor during its passage through the reactor, it seemed more like treating the symptom than the disease.
I know how commercial FCC units operate. It is my belief that none achieve the full potential of riser cracking plus closed cyclone operation. Most refiners have been happy to have the benefits of high temperature riser cracking and look on the thermal cracking as a burden of higher riser temperatures.
While quenching, either in the riser or downstream, reduces thermal cracking, significant amounts of quench material are needed. Closed cyclone operation provided the most benefit with the least burden. The benefits of closed cyclones were significant, but the full potential was usually not seen because some of the improvement was masked by the significant increase in thermal cracking of stripper vapor.
I wanted a better way to reduce thermal cracking of stripper vapor in an FCC unit using a closed cyclone system on the riser outlet. I accepted that a stripper was essential for this type of FCC operation, but did not, however, believe this material had to stagnate and crack in the reactor vessel above the stripper. Vapor had to be with the catalyst flowing down the primary cyclone diplegs into the stripper, but this vapor (as well as stripping steam) did not have to pass through the vessel volume above the stripper. All the stripper vapor did in this volume was thermally crack, and I wanted to get it out, but in a way which was compatible with the unit as it stood. I wanted a complete, rather than a partial solution to the problem of thermal cracking of stripper vapor in the reactor vessel.
Refiners wanting to reduce thermal cracking in the reactor void volume had few good options. It was possible to quench, either the entire contents of the riser reactor, or just the stripper vapor, but this had drawbacks. It would be possible to reduce thermal cracking by reducing the reactor volume, but this would involve an exorbitant capital expense. It was possible to use a stripper cap but a difficult and troublesome installation was involved. The stripper cap was a good, approach, but only a partial solution. The cap had holes in it which permitted passage of some stripper vapor through the reactor vessel volume.
I wanted to remove stripper vapor promptly to reduce thermal cracking, but did not want to have to physically isolate the stripper from reactor vessel holding the riser cyclones and riser outlet. I realized it was possible to rapidly and effectively remove stripper vapor from above the stripper using fluid dynamics rather than a cap full of holes to isolate the stripper. The simplest implementation was leave the cap off and simply use a snorkel or vapor flow tube extending from slightly above the stripper to some portion of the closed cyclone system.
Preferably, the cyclones are entirely closed, with no opening into the cyclone system except for the riser outlet and my snorkle. Closing the annular cyclone openings used in the prior art turns flow in the reactor vessel upside down. Instead of hydrocarbons rising in the reactor vessel and being thermally cracked, there is vapor phase downflow in the reactor vessel. The dome steam, preferably augmented by modest amounts of additional steam, continually forces vapor to flow down through the reactor vessel. Vapor from the riser has only two ways out of the unit—the vast majority leaves rapidly via cyclone vapor outlets while a minority (stripped hydrocarbons and stripping steam) leaves via the snorkle sealingly connected to the vapor line from the primary to the secondary cyclones.
My snorkel's function was somewhat analogous to a snorkel for a submerged submarine, with the snorkel extending just above the surface of the water, supplying air to the diesel engines. My cyclone snorkel extends down, rather than up, but its inlet should terminate just above the surface, the top level of fluidized catalyst in the FCC catalyst stripper.
I even discovered a way to support the snorkle—simply affix it to the dipleg of one or more of the primary cyclones. My snorkle is small, and handles a relatively small stream, a small amount of vapor (relative to the vapor stream discharged from the riser) with a modest amount of entrained catalyst.
Using the cyclone dipleg to support the snorkle allows the snorkle to be placed where it is needed, and supported without significantly adding “torque”. There will be little or nor problem due to differential thermal expansion, the temperature of the vapor passing up the snorkle will always be about the same as the temperature of the spent catalyst phase being discharged down the primary cyclone dipleg. Ideally, the snorkle is axially aligned with the cyclone dipleg, and discharges up into the primary cyclone vapor outlet. This will add a modest amount of solids (entrained with the stripper vapor) to the vapor phase discharged from the primary cyclone, but the secondary cyclone is equipped to handle such modest amounts of solids. The snorkle may also be within the primary cyclone dipleg, attached to a sidewall thereof, or even attached to the outside of the primary cyclone dipleg for much or all of its vertical travel.
The closed cyclone snorkel provides a way for most of the stripper vapor to be removed quickly from the reactor, without passing through the large void volume of the vessel containing the riser outlet and the stripper inlet. There will still be some hydrocarbon in the void volume. The void volume will be adequately purged using the existing dome steam injection required to prevent dome coke. If desired, thermal cracking may be even further reduced by increasing the amount of dome steam, to quench to some extent the hydrocarbons present in the reactor void volume, or additional steam may be added at different elevations. I prefer to minimize this type of steam addition, and my process will work with 500 to 5000 lbs/hr of “dome steam.”
In addition to discovering a simpler way to isolate stripper vapor from the reactor vessel, I discovered a mechanically superior type of snorkel arrangement which facilitates installation of the device.